Artificial Intelligence based Solutions for Cooperative Mobile Robots

Abstract

This paper presents the way some Artificial Intelligence techniques can contribute to obtaining an efficient solution for a cooperative robotic problem. Based on certain abstraction and sensing procedures the problem specification, the robots environment and their motion capabilities can be transposed to a finite state representation. The distributed nature of the considered application involving two robots and the reduced computation resources of individual robots conducted us to attaching robot deliberative components in an agent based implementation using a specialized software platform, namely JACK. The obtained multiagent system creates a framework for an approach that allows the interleaving of planning and execution. The agents apply an A* type bidirectional search to find the movement plan. The developed coordination protocol permits correct path generation even when the environment is changing while the robots are moving. The computational complexity of the proposed approach is low, and the system operation is supported by simulation experiments.